Compute the hamming loss. When there is only one category, this measure the accuracy.
get_hamming_loss(labels, predictions)
labels | list of labels |
---|---|
predictions | list returned by the predict command (including both the probability and the categories) |
a scalar
with the loss
library(fastrtext) data("test_sentences") model_test_path <- system.file("extdata", "model_classification_test.bin", package = "fastrtext") model <- load_model(model_test_path) sentences <- test_sentences[, "text"] test_labels <- test_sentences[, "class.text"] predictions <- predict(model, sentences) get_hamming_loss(as.list(test_labels), predictions)#> [1] 0.775